RE: [EXT] Re: [PATCH 1/2] arm64: dts: ls1088a: add more thermal zone support

From: Andy Tang
Date: Wed Jul 01 2020 - 23:34:07 EST


Hi Shawn,

Do you think it is ok if the cooling devices are shared between thermal zones?

BR,
Andy

> -----Original Message-----
> From: Amit Kucheria <amit.kucheria@xxxxxxxxxx>
> Sent: 2020å6æ30æ 14:47
> To: Andy Tang <andy.tang@xxxxxxx>
> Cc: Shawn Guo <shawnguo@xxxxxxxxxx>; Leo Li <leoyang.li@xxxxxxx>; Rob
> Herring <robh+dt@xxxxxxxxxx>; lakml <linux-arm-kernel@xxxxxxxxxxxxxxxxxxx>;
> open list:OPEN FIRMWARE AND FLATTENED DEVICE TREE BINDINGS
> <devicetree@xxxxxxxxxxxxxxx>; LKML <linux-kernel@xxxxxxxxxxxxxxx>
> Subject: Re: [EXT] Re: [PATCH 1/2] arm64: dts: ls1088a: add more thermal
> zone support
>
> Caution: EXT Email
>
> On Tue, Jun 30, 2020 at 12:07 PM Andy Tang <andy.tang@xxxxxxx> wrote:
> >
> >
> >
> > > -----Original Message-----
> > > From: Amit Kucheria <amit.kucheria@xxxxxxxxxx>
> > > Sent: 2020å6æ30æ 13:37
> > > To: Andy Tang <andy.tang@xxxxxxx>
> > > Cc: Shawn Guo <shawnguo@xxxxxxxxxx>; Leo Li <leoyang.li@xxxxxxx>;
> > > Rob Herring <robh+dt@xxxxxxxxxx>; lakml
> > > <linux-arm-kernel@xxxxxxxxxxxxxxxxxxx>;
> > > open list:OPEN FIRMWARE AND FLATTENED DEVICE TREE BINDINGS
> > > <devicetree@xxxxxxxxxxxxxxx>; LKML <linux-kernel@xxxxxxxxxxxxxxx>
> > > Subject: Re: [EXT] Re: [PATCH 1/2] arm64: dts: ls1088a: add more
> > > thermal zone support
> > >
> > > Caution: EXT Email
> > >
> > > On Tue, Jun 30, 2020 at 10:58 AM Andy Tang <andy.tang@xxxxxxx>
> wrote:
> > > >
> > > >
> > > >
> > > > > -----Original Message-----
> > > > > From: Amit Kucheria <amit.kucheria@xxxxxxxxxx>
> > > > > Sent: 2020å6æ30æ 13:12
> > > > > To: Andy Tang <andy.tang@xxxxxxx>
> > > > > Cc: Shawn Guo <shawnguo@xxxxxxxxxx>; Leo Li
> > > > > <leoyang.li@xxxxxxx>; Rob Herring <robh+dt@xxxxxxxxxx>; lakml
> > > > > <linux-arm-kernel@xxxxxxxxxxxxxxxxxxx>;
> > > > > open list:OPEN FIRMWARE AND FLATTENED DEVICE TREE BINDINGS
> > > > > <devicetree@xxxxxxxxxxxxxxx>; LKML
> > > > > <linux-kernel@xxxxxxxxxxxxxxx>
> > > > > Subject: [EXT] Re: [PATCH 1/2] arm64: dts: ls1088a: add more
> > > > > thermal zone support
> > > > >
> > > > > Caution: EXT Email
> > > > >
> > > > > On Tue, Jun 30, 2020 at 8:56 AM <andy.tang@xxxxxxx> wrote:
> > > > > >
> > > > > > From: Yuantian Tang <andy.tang@xxxxxxx>
> > > > > >
> > > > > > There are 2 thermal zones in ls1088a soc. Add the other
> > > > > > thermal zone node to enable it.
> > > > > > Also update the values in calibration table to make the
> > > > > > temperatures monitored more precise.
> > > > > >
> > > > > > Signed-off-by: Yuantian Tang <andy.tang@xxxxxxx>
> > > > > > ---
> > > > > > .../arm64/boot/dts/freescale/fsl-ls1088a.dtsi | 100
> > > > > > +++++++++++-------
> > > > > > 1 file changed, 62 insertions(+), 38 deletions(-)
> > > > > >
> > > > > > diff --git a/arch/arm64/boot/dts/freescale/fsl-ls1088a.dtsi
> > > > > > b/arch/arm64/boot/dts/freescale/fsl-ls1088a.dtsi
> > > > > > index 36a799554620..ccbbc23e6c85 100644
> > > > > > --- a/arch/arm64/boot/dts/freescale/fsl-ls1088a.dtsi
> > > > > > +++ b/arch/arm64/boot/dts/freescale/fsl-ls1088a.dtsi
> > > > > > @@ -129,19 +129,19 @@
> > > > > > };
> > > > > >
> > > > > > thermal-zones {
> > > > > > - cpu_thermal: cpu-thermal {
> > > > > > + core-cluster {
> > > > > > polling-delay-passive = <1000>;
> > > > > > polling-delay = <5000>;
> > > > > > thermal-sensors = <&tmu 0>;
> > > > > >
> > > > > > trips {
> > > > > > - cpu_alert: cpu-alert {
> > > > > > + core_cluster_alert:
> > > > > core-cluster-alert
> > > > > > + {
> > > > > > temperature =
> > > <85000>;
> > > > > > hysteresis =
> <2000>;
> > > > > > type = "passive";
> > > > > > };
> > > > > >
> > > > > > - cpu_crit: cpu-crit {
> > > > > > + core_cluster_crit:
> > > > > > + core-cluster-crit {
> > > > > > temperature =
> > > <95000>;
> > > > > > hysteresis =
> <2000>;
> > > > > > type = "critical";
> @@
> > > > > -150,7
> > > > > > +150,7 @@
> > > > > >
> > > > > > cooling-maps {
> > > > > > map0 {
> > > > > > - trip =
> <&cpu_alert>;
> > > > > > + trip =
> > > > > <&core_cluster_alert>;
> > > > > > cooling-device =
> > > > > > <&cpu0
> > > > > THERMAL_NO_LIMIT THERMAL_NO_LIMIT>,
> > > > > > <&cpu1
> > > > > > THERMAL_NO_LIMIT THERMAL_NO_LIMIT>, @@ -163,6 +163,26
> @@
> > > > > > };
> > > > > > };
> > > > > > };
> > > > > > +
> > > > > > + soc {
> > > > > > + polling-delay-passive = <1000>;
> > > > > > + polling-delay = <5000>;
> > > > > > + thermal-sensors = <&tmu 1>;
> > > > > > +
> > > > > > + trips {
> > > > > > + soc-alert {
> > > > > > + temperature =
> > > <85000>;
> > > > > > + hysteresis =
> <2000>;
> > > > > > + type = "passive";
> > > > > > + };
> > > > > > +
> > > > > > + soc-crit {
> > > > > > + temperature =
> > > <95000>;
> > > > > > + hysteresis =
> <2000>;
> > > > > > + type = "critical";
> > > > > > + };
> > > > > > + };
> > > > > > + };
> > > > >
> > > > > You should also add a cooling-maps section for this thermal zone
> > > > > given that it has a passive trip type. Otherwise there is no use
> > > > > for a
> > > passive trip type.
> > > > It is better to have a cooling device. But there is only one
> > > > cooling device on this platform which is used by core-cluster. So
> > > > there is no extra
> > > cooling device for it.
> > > > This zone can take action when critical temp is reached. So it is still
> useful.
> > > > What do you suggest?
> > >
> > > If the action taken by the core-cluster cooling-maps is the only one
> > > that can be taken, I suggest getting rid of the the soc-alert
> > > passive trip completely. It is not of any use.
> > >
> > > If there is a chance that your soc thermal-zone can heat up before
> > > your cpu-cluster zone (unlikely), you could use the same cooling
> > > device (cpu0, cpu1) for soc thermal zone too.
> > Thanks for your suggestion.
> > I was told that a cooling-maps can not be applied to more than one thermal
> zone.
> > So the only option is to remove the alert trip.
>
> The cooling-map is unique to the thermal-zone. But the cooling devices may
> be reused.
>
> See arch/arm64/boot/dts/qcom/sdm845.dtsi for an example.
>
> So, in your case, you would have a cooling-map section that ties the soc-alert
> trip to the same cooling devices as in cpu-cluster cooling map.